Close
0%
0%

3×5:8λ PMR antenna :Clean, Matched, 3D Printable

Low-cost PMR446 collinear with high gain, accurate impedance match, and a fully 3D-printed structure for reproducible builds.

Similar projects worth following
The PMR446 band (446 MHz) is one of the few truly license-free radio communication options available in France. It allows the use of handheld radios (walkie-talkies) for short-range communication. However, this freedom comes with strict regulations: power is limited to 0.5 W ERP, antennas must be fixed and non-removable (for transmissions). In urban environments, the range rarely exceeds a few kilometers; under ideal conditions, it may reach up to 15 km. This antenna was specifically designed to improve local reception on the PMR446 band. The design is based on rigorous RF principles—phasing, matching, impedance tuning—and has been fully characterized with a NanoVNA. It is easily reproducible thanks to 3D printing. Construction is inexpensive, using readily available materials such as 4 mm brass tube and 1.5 mm² copper wire. All necessary files for 3D print are provided, enabling anyone to build a functional version.

📶 Why Stack Radiating Elements?

Stacking multiple vertical radiators along the same axis — the core principle behind collinear antennas — concentrates energy toward the horizon and flattens the vertical radiation pattern. A simple quarter-wave whip (¼ λ) produces a lobe at approximately 25° elevation with a reference gain of 0 dBd. Upgrading to a half-wave (½ λ) shifts the main lobe down to ≈ 15° and yields a gain of about +2 dBd. By chaining together three identical and correctly phased sections, typical gains reach around +6 dBd, with a main lobe just 3–5° above the horizon.

This constructive superposition of fields is what makes collinear designs attractive for increasing horizontal range.

However, this gain does not scale indefinitely. Beyond a few sections, geometric and phase alignment errors accumulate, side lobes begin to grow, and mechanical constraints or signal cancellation reduce overall efficiency. That’s why most practical collinears stop at two or three (sometimes five for commercial device) radiating sections.


Radiation Angle vs Gain (Illustrative)

ConfigurationElevation Angle (°)Gain (dBd)Notes
¼ λ monopole~25°0 dBdBasic whip, wide lobe
½ λ vertical~15°+2 dBdNarrower lobe
2 × 5/8 λ collinear~8–10°+4–5 dBdRequires phasing stub
3 × 5/8 λ collinear~3–5°+6–7 dBdFlatter lobe, compact design
≥ 4 × 5/8 λ (theoretical)<3°+7–8 dBdDiminishing returns + complex +
length


🛑📡The Non-Radiation of the Folded Stub: Physical Justification and Current Integral

In the folded stub (three parallel conductors connected by two 180° bends), the cancellation of radiation does not arise from differences in optical path: since the conductors are very close together (d ≪ λ), every point on the stub is virtually at the same distance from the far-field observer, so the geometric phases are aligned.

The real mechanism is linked to the current distribution: as you move along the stub, the current reverses direction at each change of conductor (up, down, up again). Mathematically, this is expressed by a far-field integral :


where I(z) alternates in sign along each straight segment due to the folded geometry.

As a result of this alternation (harmonic character of the current plus the inversion of its direction along the stub), the integral of the contributions from the three straight sections is nearly zero over the whole stub, and the bends only contribute negligibly : on the order of kd power 2.

In this image, the stub is visible, with the copper wire guided along the structure. On the right, a 4 mm brass tube is inserted into the PLA body. The tube is then crimped onto the copper wire using a crimping tool (if you don’t have one, a standard pair of pliers can be used instead).


In summary, the folded stub is almost electromagnetically silent, because the variation in current and the reversal of its direction within the stub enforce an almost complete cancellation of the far-field. This result is fundamental and is the same principle used in Franklin arrays and other phased-line antennas. Some commercial antenna designs also employ similar geometries for their stub.


🌀Capacitive Matching of the Radiating Element and Impedance Autotransformer

A vertical monopole antenna resonates ideally when it measures a quarter-wavelength (¼ λ): at this length, the input impedance is largely resistive and well matched (typically 36 Ω with a perfect ground plane) and has very little reactive component. However, as soon as the element is lengthened beyond λ/4—up to about 5/8 λ, as is the case here for our first section to optimize the radiation pattern—the tip of the element accumulates significant charge, creating a capacitive reactance: the feedpoint “sees” a typical impedance of about...

Read more »

stub.STEP

step - 537.45 kB - 11/08/2025 at 08:41

Download

corps.STEP

step - 3.04 MB - 11/08/2025 at 08:39

Download

mandrin.STEP

step - 537.47 kB - 11/08/2025 at 08:39

Download

bouchon.STEP

step - 159.75 kB - 11/08/2025 at 08:39

Download

corps.png

Portable Network Graphics (PNG) - 262.57 kB - 05/20/2025 at 16:49

Preview

View all 19 files

  • 1 × SO-239 connector RF feedpoint, chassis mount
  • 1 × 4 mm brass tube (hollow) allows copper wire to be inserted inside
  • 1 × 3 mm brass tube horizontal radiator elements
  • 1 × 1.5 mm^2 copper wire tuning coil & stubs
  • 1 × PLA 1.75mm diameter

  • RF losses in PVC tube

    Bertrand Selva08/10/2025 at 06:57 0 comments

    Here’s a estimate of RF losses in PVC tube.

    Assumptions: D = 20 mm, wall = 1 mm → r2 = 10 mm, r1 = 9 mm; centered radiator a ≈ 1 mm; f = 446 MHz so λ ≈ 0.672 m and R ≈ λ/(2*pi) ≈ 107 mm; epsilon_r ≈ 3.2; tan(delta) ≈ 0.01–0.02.

    Fraction of energy in the PVC wall

    Formula (thin-wire in a coax-like shell): Wd/Wtot = (epsilon_r * ln(r2/r1)) / (epsilon_r*ln(r2/r1) + ln(r1/a) + ln(R/r2))

    Numbers (mm): ln(r2/r1) = ln(10/9) = 0.10536 epsilon_rln(r2/r1) = 3.20.10536 = 0.337 ln(r1/a) = ln(9/1) = 2.197 ln(R/r2) = ln(107/10) = 2.370

    So: Wd/Wtot = 0.337 / (0.337 + 2.197 + 2.370) = 0.337 / 4.905 = 0.069 ≈ 6.9>#/p###

    Dielectric loss (referred to radiated power)

    Formula: Pdiel/Prad ≈ Qrad * tan(delta) * (Wd/Wtot)

    Typical dry case: Qrad = 10 (thin wire), tan(delta) = 0.01 → Pdiel/Prad = 100.010.069 = 0.0069 (0.69%) → ≈ 0.03 dB

    Pessimistic dry case: Qrad = 20, tan(delta) = 0.02 → Pdiel/Prad = 200.020.069 = 0.0276 (2.76%) → ≈ 0.12 dB

    Conclusion (D = 20 mm, wall 1 mm, dry, retuned): additional RF loss ≈ 0.03–0.12 dB, typically ~0.05–0.08 dB. The dominant effect remains detuning but as I explained I correct during tuning (and it's true the effect of PVC is important).

View project log

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates